Obstacle detection for autonomous vehicles

The academic society has for a long time had a great interest in building autonomous vehicles. However there are not that many successful commercial applications. One reason for that is probably that the obstacle detection has been too expensive or did not work satisfactory. Recently there have been some new interesting low cost systems launched. Those are based on scanning the environment to get a distance matrix. This matrix can be used to analyze what you have ahead and steer the vehicle around the obstacles.

In this thesis work you will design an obstacle detection system for an autonomous vehicle. You will use a low cost environmental scanning sensor to analyze where the obstacles are and then make a control algorithm to avoid the obstacles. The aim is that the vehicle shall pass around the obstacle without touching it. The control of the vehicle is done in an embedded system based on an ARM processor. In the processor embedded Linux is used as OS. You will focus only on the control algorithm; all other drivers and functions already exist.

Below is a summary of the thesis work content.
- Literature study on obstacle detection for autonomous vehicles.
- Make analysis of different system set-ups regarding cost, robustness and performance.
- Chose one system set-up and purchase the needed HW.
- Design algorithms for the obstacle detection and how to avoid them.
- Implement the algorithms in an ARM processor with embedded Linux.
- Verify the control algorithms in a real environment.
- Write final report.

Key competence
- Design of embedded systems
- Image processing
- Control engineering
- C/C++ programming
- SW design in a real time OS environment
- Embedded Linux
- Basic electronics design